Daily Papers of 2026-09-02

  1. StudentSim: Training LLM-based Student Simulators 485 upvotes, #1 of 2026-09-02
  2. Qwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving 378 upvotes, #2 of 2026-09-02
  3. SMELT: Scaling Laws for Compute-Matched MoE Looped Transformers 99 upvotes, #3 of 2026-09-02
  4. UI-Venus-2 Technical Report 63 upvotes, #4 of 2026-09-02
  5. ZimaBlue: Evolving Generalizable World Action Models through Scalable Video Pre-training 51 upvotes, #5 of 2026-09-02
  6. H3-World: Turning Language Understanding into World Control 50 upvotes, #6 of 2026-09-02
  7. Hi-Q: Hierarchical Evidence-guided Query Refinement for Multi-Hop Question Answering 37 upvotes, #7 of 2026-09-02
  8. From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix 35 upvotes, #8 of 2026-09-02
  9. Evaluating Multimodal LLMs as Generalist Vision-Language-Action Agents for Drone Control: Commanding, Approaching, Tracking and Searching 27 upvotes, #9 of 2026-09-02
  10. Uncovering Understanding-Generation Synergy in Native Unified Multimodal Models: From Representation, Task to System 23 upvotes, #10 of 2026-09-02
  11. DiagEvo: Diagnosis-Guided Self-Evolution via Hierarchical Error Memory 22 upvotes, #11 of 2026-09-02
  12. AgentJudgeBench: A Multi-Difficulty Benchmark for Evaluating LLM Judges on Agentic Tool-Calling 20 upvotes, #12 of 2026-09-02
  13. Safin-1: Safety from Within through Memory-Native State Evolution 20 upvotes, #12 of 2026-09-02
  14. E-Commerce Bench: Evaluating LLM Agents on Long-Horizon Autonomous Business Operation 17 upvotes, #14 of 2026-09-02
  15. Harness-of-Harness: Multi-Day Autonomous Software Development with Continual Improvement 17 upvotes, #14 of 2026-09-02
  16. EM^2Mem: Event-Centric Multimodal Memory for Large Language Models 14 upvotes, #16 of 2026-09-02
  17. Agents in the Large: Perception-Centered Architecture for Persistent Agents 11 upvotes, #17 of 2026-09-02
  18. InternReviewer & InternAdvocate: Objective Reward and Evaluation for Agentic Reinforcement Learning in Peer Review and Rebuttal 9 upvotes, #18 of 2026-09-02
  19. Learning Where Outcomes Change:Credit-Addressable Reasoning for Multimodal Geometry 9 upvotes, #18 of 2026-09-02
  20. Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs 9 upvotes, #18 of 2026-09-02
  21. Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall 8 upvotes, #21 of 2026-09-02
  22. The Mechanics of Democratic Dominance: A System Dynamics Paradigm for Dynamic Consent Engineering 7 upvotes, #22 of 2026-09-02
  23. Recursive Criticality of AI Self-Improvement 7 upvotes, #22 of 2026-09-02
  24. Adapting Without Gradients: Affine Statistics Transport and What Its Certificate Can Tell You 7 upvotes, #22 of 2026-09-02
  25. Agent Memory Is a Surface for Endogenous Authorization Laundering 7 upvotes, #22 of 2026-09-02
  26. DramaChain Bench: An End-to-End Benchmark for Short-Drama Generation 6 upvotes, #26 of 2026-09-02
  27. Token-Efficient Data Reasoning Agents via Adaptive Structuring of Unstructured Data 5 upvotes, #27 of 2026-09-02
  28. ReFlowSET: Representation-Aligned Latent Flow Matching for SAR-to-EO Image Translation 5 upvotes, #27 of 2026-09-02
  29. Does Imitation Learning Preserve Temporal Robustness in Dexterous Manipulation? An Expert-Learner Comparison Across Task Execution Speeds 3 upvotes, #29 of 2026-09-02

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.